- Email: [email protected]
Correlation between early clinical response after catheter removal and diagnosis of catheter-related bloodstream infection
Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457 www.elsevier.com/locate/diagmicrobio
Clinical Study
Correlation between early clinical response after catheter removal and diagnosis of catheter-related bloodstream infection Bassam Shukrallah, Hend Hanna, Ray Hachem, Dany Ghannam, Ioannis Chatzinikolaou, Issam Raad4 The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA Received 22 November 2006; accepted 16 March 2007
Abstract We conducted a retrospective post hoc analysis of prospectively collected data of cancer patients with central venous catheters (CVCs) who developed bacteremia with positive quantitative blood cultures (QBCs) drawn simultaneously through peripheral vein and CVC and which grew the same microorganisms from both blood cultures. We investigated whether clinical response of bacteremia, within 24, 48, or 72 h post-CVC removal, could be diagnostic of catheter-related bloodstream infection (CRBSI) when compared with microbiologic methods. Clinical response to antimicrobial therapy within 24 h of CVC removal in a patient with bacteremia was found to be highly suggestive of CRBSI, a finding that correlated well with semiquantitative catheter cultures and differential QBCs. However, response to antimicrobial therapy at z 48 h after CVC removal was less likely to be diagnostic of CRBSI and could reflect a response to antimicrobial therapy irrespective of the source of the bloodstream infections. D 2007 Elsevier Inc. All rights reserved. Keywords: Catheter removal; Bloodstream infections; Diagnosis of infections
1. Introduction The high frequency of catheter-related bloodstream infection (CRBSI), between 200 and 400 000 episodes occurring annually in the United States, is a public heath challenge ([Raad, 1998; Maki and Mermel, 1998; Raad and Hanna, 2002]). Not only do these infections present a significant monetary burden on the current healthcare system, where the estimated cost of managing each episode can range from $34 000 to $56 000 (Rello et al., 2000; Pittet et al., 1994), but they also are associated with a mortality rate of 12% to 35% (Collingnon, 1994; Pittet et al., 1994; Klugger and Maki, 1999). Accurately diagnosing CRBSI is essential for the successful management of such infections. Historically, clinicians arriving at the right diagnosis relied
4 Corresponding author. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA. Tel.: +1-713-792-7943; fax: +1713-792-8233. E-mail address: [email protected] (I. Raad). 0732-8893/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2007.03.012
on catheter removal and culture. However, only 4% to 10% of central venous catheters (CVCs) removed due to a suspected CRBSI were found to be colonized (Widmer et al., 1992). To address this problem, we developed newer diagnostic methods, such as differential time to positivity (DTP) and quantitative blood cultures (QBC) (Raad et al., 2004; Chatzinikolaou et al., 2006). Both of these methods provide sensitive and specific diagnostic capabilities without requiring the removal of the catheter. The Centers for Disease Control and Prevention (CDC) and Infectious Diseases Society of America (IDSA) recently incorporated these tests into their guidelines for the prevention of intravascular catheter-related infections (Mermel et al., 2001; O’Grady et al., 2002). Clinical response, defined as resolution of fever and other clinical manifestations of infection after removal of CVC, has also been considered in the absence of laboratory confirmation, as an indirect evidence of CRBSI (Pearson, 1996). This method has not been studied well, nor has it been compared with the established microbiologic methods. Furthermore, it lacks temporal parameters in the literature. The purpose of this study was to determine whether clinical
454
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457
Table 1 Clinical characteristics of patients with CRBSI and non-CRBSI Characteristic, n (%), unless indicated otherwise
CRBSI, n = 70
Non-CRBSI, n = 43
P
Sex, male Mean age F SD (years) Underlying malignancy Leukemia Lymphoma/myeloma Solid tumor ICU CVC type Silicone Rigid wall Port Quinton Neutropenia ( b 500 cells/mm3) Thrombocytopenia ( b 100 103/AL)
45 (64.3) 51.6 F 15.6
30 (69.8) 53.0 F 14.4
.55 .65
11 (15.7) 16 (22.9) 43 (61.4) 14 (20.0)
19 8 16 7
(44.2) (18.6) (37.2) (16.3)
.001 .59 .01 .62
62 3 4 1 22
(88.6) (4.3) (5.7) (1.4) (31.4)
29 (67.4) 11 (25.6) 2 (4.7) 1 (2.3) 26 (60.5)
.006 .001 .59 .62 .002
37 (52.9)
30 (69.8)
.08
response, within 24, 48, or 72 h of CVC removal, is indeed a useful diagnostic tool for CRBSI. Our current study focused on whether clinical response within 24 to 72 h, post-catheter removal, offers any diagnostic applicability when compared with semiquantitative catheter cultures (SQCCs) and QBCs. 2. Methods Between September 1999 and November 2000, we prospectively identified 113 cancer patients who, while having CVC in place, developed fever and had positive quantitative cultures of blood that was drawn simultaneously through peripheral vein and through the CVC and which grew the same microorganisms from both blood cultures. Patients were included in the study if their catheters were removed and cultured and if they did not have a complicated or disseminated bloodstream infection. 2.1. Definitions Catheter-related bloodstream infection. This is defined as the 1) fever and other clinical manifestations of bloodstream infection in a patient with CVC and 2) the presence of positive simultaneous QBC from the CVC and peripheral vein yielding the same organism, with 3) no apparent source for the bloodstream infection other than the catheter and 4) the presence of SQCC with significant catheter tip colonization with z 15 CFU of the same organism (same species and antibiogram) isolated from the blood cultures, and/or simultaneous QBC whereby the number of colonies isolated from the blood drawn through the CVC are at least 5-fold greater ( z 5:1) than the number of colonies isolated from blood drawn via a peripheral vein. CVC colonization. Significant colonization of the catheter tip was defined as a positive SQCC using the roll plate
method whereby at least 15 colonies of an organism was cultured from the catheter tip. Clinical response. This is the resolution of fever and other clinical manifestations of infection. 2.2. Culture techniques 2.2.1. Blood cultures After antiseptic cleansing of the skin and catheter hub, simultaneous quantitative and qualitative blood cultures were performed on blood drawn through the peripheral vein and the CVC’s hub. Blood samples drawn for cultures were handled as follows: 7 to 10 mL of blood was drawn through the CVC and discarded to avoid mixing with antibiotics or other agents that could have antimicrobial activity and that may have been given previously through the CVC. Subsequently, 20 mL of blood were drawn through the CVC and divided into 2 portions. Ten milliliters was placed in isolator tubes (Isolator 10; Wampole, Cranbury, NJ) to be cultured quantitatively according to the lysis centrifugation method as previously described (Dorn et al., 1979). Another 10 mL of blood was placed in a regular aerobic blood culture bottle (Aerobic 26+; BD Diagnostics, Sparks, MD). 2.2.2. Catheter culture Catheters were removed aseptically, and a 5-cm segment of the removed catheter tip was cultured using the semiquantitative roll plate method (Maki et al., 1977). 2.3. Statistical analysis Patients with CRBSI were compared with those with non-CRBSIs. The significance of the differences between the 2 groups was determined through the v 2 test or the Fisher’s exact test, as appropriate for categoric variables.
Table 2 Microbial organisms causing bloodstream infections in 113 cancer patients Organismsa
CRBSI (n = 70)
Non-CRBSI (n = 43)
P
Gram-positive organisms Coagulase-negative staphylococci Staphylococcus aureus a-Hemolytic Streptococcus Enterococcus spp. Other Gram-positive organismsb Gram-negative organisms Pseudomonas aeruginosa Escherichia coli Klebsiella pneumoniae Other Gram-negative organismsc Candida spp. Acid fast bacilli
51 32 9 3 2 5 13 3 2 5 3 5 1
30 (69.8%) 10 (23.3%) 12 (27.9%) 5 (11.6%) 2 (4.7%) 1 (2.3%) 11 (25.6%) 2 (4.7%) 4 (9.3%) 1 (2.3%) 4 (9.3%) 2 (4.7%) 0
.83 .018 .079 .26 .63 .4 .48 N .99 .2 .4 .42 .71 N .99
(73%) (45.7%) (12.9%) (4.3%) (2.9%) (7.1%) (19%) (4.3%) (2.9%) (7.1%) (4.3%) (7.1%) (1.4%)
a Numbers in parentheses are percentages of the total number of each isolated species. b h-Hemolytic Streptococcus, Bacillus spp., Corynebacterium, Micrococcus, and Stomatococcus. c Acinetobacter, Alcaligenes, Citrobacter, Enterobacter, Flavimonas, and Stenotrophomonas.
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457 Table 3 Clinical response after catheter removal Clinical response
CRBSI (n = 70)
Non-CRBSI (n = 43)
Within 24 h of CVC removal, n (%) Within 48 h of CVC removal, n (%) Within 72 h of CVC removal, n (%)
27 (37.1) 31 (42.9) 34 (47.1)
9 (23.3) 15 (34.9) 16 (37.2)
P .05 .323 .238
Student’s t test or Mann–Whitney U test was used for continuous variables. All P values were based on 2-tailed tests of significance with a level of significance at P V 0.05. Statistical analyses were performed using the statistical computing package SPSS (version 12.0 for Windows; SPSS, Chicago, IL). 3. Results Between September 1999 and November 2000, there were a total of 191 patients who had CVCs, developed fever, and had QBC done. The study included 113 of 191 patients who had their CVCs removed and cultured and who did not have complicated or disseminated infections. According to the definitions used in the study, there were 70 patients who had CRBSI and 43 who had non-CRBSI. The 2 groups were comparable in age, sex, and intensive care unit (ICU) admission status. Patients with CRBSI had solid tumor more often as their underlying malignancy ( P = 0.012), were less often neutropenic ( P = 0.002) and tended to be less thrombocytopenic ( P = 0.08), and had silicone CVCs more often (Table 1). Table 2 shows the distribution of microbial organisms causing CRBSI and non-CRBSI. The 2 groups were comparable in the distribution of major categories of infecting organisms (Gram-positive organisms, Gram-negative organisms, Candida spp., and acid fast bacilli). However, the frequency of coagulase-negative staphylococci causing CRBSI was significantly higher than those causing non-CRBSI (45.7% versus 23.3%, P = 0.016). Patients with CRBSI responded more often within the first 24 h after CVC removal ( P = 0.05) than did patients with non-CRBSI. This difference between the responses of the 2 groups disappeared at 48 and at 72 h post-CVC removal, when clinical response (CR) in patients with CRBSI was comparable to that of patients with non-CRBSI (Table 3). 4. Discussion This study shows that clinical response to antimicrobial therapy occurring within 24 h after the removal of the CVC in a patient with bloodstream infection may be suggestive of CRBSI. This suggested that clinical diagnostic criterion correlates well with microbiologic methods involving SQCCs and differential QBCs. However, response to antimicrobial therapy within 48 h or longer after CVC removal is less likely to be diagnostic of CRBSI and could
455
reflect the activity of the antimicrobial agents against the microorganism causing the bloodstream infections, irrespective of its source. The clinical diagnosis of CRBSI has been a challenge. On their own, clinical signs and symptoms are not useful in establishing the diagnosis of CRBSI because of their limited sensitivity or specificity (Mermel et al., 2001). The occurrence of fever or positive blood cultures in a patient with an indwelling CVC might not necessarily suggest, on its own, that the catheter is the source of the bloodstream infection. On the other hand, purulent discharge or inflammation involving the catheter exit site, the tunnel, or the pocket of the port might suggest that the catheter is a source of the bloodstream infection. However, although such findings are highly specific, they are not always associated with isolation of organisms that commonly cause CRBSI such as Staphylococcus epidermidis or Candida spp. Furthermore, evaluation of suspected infection in the presence of acute neutropenia is especially difficult because typical signs of inflammation may be markedly masked by the lack of neutrophils; hence, it is less likely that these patients would develop purulence at an infected catheter exit site. Because clinical findings are often associated with either poor sensitivity or poor specificity in establishing the diagnosis of CRBSI, the IDSA has recommended establishing the diagnosis based on microbiologic methods that include either quantitative catheter cultures or other differential blood culture methods, such as differential QBC or DTP (Raad et al., 2004; Chatzinikolaou et al., 2006; Mermel et al., 2001). Several studies have supported the use of such microbiologic methods (Raad et al., 2004; Chatzinikolaou et al., 2006; Flynn et al., 1988; Safdar et al., 2005; Blot et al., 1999; Gaur et al., 2003). In a recent metaanalysis, differential QBCs used in establishing the diagnosis of CRBSI in the study were considered the most highly specific and sensitive methods in making the diagnosis of such infection (Safdar et al., 2005). However, differential QBCs are not readily available in most hospitals in the United States or worldwide, and they are labor intensive and expensive. Furthermore, semiquantitative and quantitative catheter cultures are also costly and require the removal of the catheter. In their 1996 guidelines on the prevention of intravascular catheter-related infections, the CDC suggested that, in the absence of laboratory confirmation, defervescence after the removal of the suspected CVC from a patient with a bloodstream infection may be considered an indirect evidence of CRBSI (Pearson, 1996). Other investigators and reviewers have also relied on such clinical criteria (Blot et al., 1999). However, such criteria have not been well defined or tested against strict microbiologic methods that define CRBSI. Furthermore, neither the CDC nor the other investigators who referred to such clinical criteria outlined the time parameters necessary in establishing the diagnosis of CRBSI (24, 48, 72, or 96 h after CVC removal).
456
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457
The data in this current study suggest that a clinical response represented by defervescence after early CVC removal (within 24 h of such intervention) in patients with a bloodstream infection who are on appropriate antimicrobial therapy is suggestive of CRBSI. In our study, such early response does correlate well with the microbiologic definition of CRBSI, based on SQCCs and differential QBC as defined by the IDSA (Mermel et al., 2001). However, a clinical response occurring after more than 24 h of CVC removal does not suggest the diagnosis of CRBSI and does not correlate with the well-defined microbiologic criteria of CRBSI. The fact that the early response is indicative of the microbiologic diagnosis of CRBSI, rather than a later response, in a patient population on an appropriate antimicrobial therapy is quite revealing. Patients with uncomplicated bloodstream infections usually defervesce after at least 48 to 72 h of appropriate antimicrobial therapy (Fowler et al., 2003), irrespective of the source of the infection (catheter or otherwise). This occurs in cases of uncomplicated bloodstream infection in the absence of a deep-seated infection such as endocarditis or septic thrombosis. Patients with a definite uncomplicated CRBSI are more likely to respond earlier if the catheter source of the bloodstream infection is removed. Hence, the first 24 h are crucial in making the distinction between CRBSI and non-CRBSI cases. After 24 h, and particularly after 48 h of antimicrobial therapy, all patients with uncomplicated bloodstream infections would equally respond. This study, however, has several limitations. One of the major limitations is the small number of patients involved that included 113 patients who received appropriate antimicrobial therapy with an agent that is active against the organism isolated from the blood. The impact of CVC removal within 24 h was evaluated in a total of 36 patients, and the impact within 48 h was evaluated in only 48 patients. Hence, the lack of a difference, particularly in patients who defervesced within 48 or 72 h of CVC removal, could be related to the small numbers and the lack of power rather than the absence of impact of the CVC removal on establishing the diagnosis. Furthermore, the design of the study in terms of being a retrospective post hoc analysis of a prospective study represents another limitation. Finally, the fact that this study included cancer patients with different underlying malignancy and level of neutropenia might have affected the frequency of the early response in the CRBSI group. Patients with CRBSI were less likely to be neutropenic or to have underlying hematologic malignancy and, hence, they were more likely to respond earlier to antimicrobial therapy than the more immunocompromised neutropenic patients with underlying hematologic malignancy (Table 1). Finally, this method of establishing the clinical diagnosis requires the removal of the CVC; hence, in this current study, the impact of this method in establishing the diagnosis might have been limited to those cases of CRBSI
where the catheter was highly suspected and removed early during the bacteremia. Therefore, this method of clinically diagnosing the CRBSI might be useful retrospectively in surveillance rather than in establishing the diagnosis at the bedside. In addition, because this method requires the removal of the CVC, relying on it might lead to unnecessary and wasteful removal of the important intravascular device in the absence of other vascular access sites. In conclusion, based on a small number of patients, our data suggest that early clinical response to antimicrobial therapy, characterized by defervescence after 48 h of CVC removal, does not provide direct or indirect evidence suggesting the diagnosis of CRBSI. However, in a subset of patients who defervesce early within 24 h after the CVC is removed, the diagnosis of CRBSI is more likely and should be considered.
References Blot F, et al (1999) Diagnosis of catheter-related bacteraemia: a prospective comparison of the time to positivity of hub-blood versus peripheralblood cultures. Lancet 354:1071 – 1077. Chatzinikolaou I, et al (2006) Differential quantitative blood cultures for the diagnosis of catheter related bloodstream infections associated with short and long-term catheters: a prospective study. Diagn Microbiol Infect Dis 44:1834 – 1835. Collingnon PJ (1994) Intravascular catheter associated sepsis: a common problem. The Australian study on intravascular catheter associated sepsis. Med J Aust 161:374 – 378. Dorn GL, et al (1979) Improved blood culture technique based on centrifugation: clinical evaluation. J Clin Microbiol 9:391 – 396. Flynn PM, et al (1988) Differential quantitation with a commercial blood culture tube for diagnosis of catheter-related infection. J Clin Microbiol 26:1045 – 1046. Fowlerr Jr VG, et al (2003) Clinical identifiers of complicated Staphylococcus aureus bacteremia. Arch Intern Med 163:2066 – 2071. Gaur AH, et al (2003) Difference in time to detection: a simple method to differentiate catheter-related from non–catheter-related bloodstream infection in immunocompromised pediatric patients. Clin Infect Dis 37:469 – 475. Klugger DM, Maki DG (1999) The relative risk of intravascular device related infections in adults [Abstract]. Abstracts of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Francisco (CA)7 American Society for Microbiology, pp 514. Maki DG, et al (1977) A semiquantitative culture method for identifying intravenous-catheter–related infection. N Engl J Med 296:1305 – 1309. Maki DG, Mermel LA (1998) Infections due to infusion therapy. In Hospital infections. 4th ed. Eds, JV Bennett and PS Brachman. Philadelphia7 Lippincott-Raven, pp 689 – 724. Mermel LA, et al (2001) Guidelines for the management of intravascular catheter-related infections. Clin Infect Dis 32:1249 – 1272. O’Grady NP, et al (2002) Guidelines for the prevention of intravascular catheter-related infections. MMWR Recomm Rep 51:1 – 29. Pearson ML (1996) Guideline for prevention of intravascular device-related infections. Part I: Intravascular device-related infections: an overview. The Hospital Infection Control Practices Advisory Committee. Am J Infect Control 24:262 – 277. Pittet D, et al (1994) Nosocomial bloodstream infections in critically ill patients. Excess length of stay, extra costs, attributable mortality. JAMA 271:1598 – 1601.
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457 Raad I (1998) Intravascular-catheter–related infections. Lancet 351: 893 – 898. Raad II, Hanna HA (2002) Intravascular catheter-related infections. New horizons and recent advances. Arch Intern Med 162:871 – 878. Raad I, et al (2004) Differential time to positivity: a useful method for diagnosing catheter-related bloodstream infections. Ann Intern Med 140:18 – 25.
457
Rello J, et al (2000) Evaluation of outcome of the intravenous catheterrelated infections in critically ill patients. Am J Respir Crit Care Med 162:1027 – 1030. Safdar N, et al (2005) Meta-analysis: methods for diagnosing intravascular device-related bloodstream infection. Ann Intern Med 42:451 – 466. Widmer AF, et al (1992) The clinical impact of culturing central venous catheters. A prospective study. Arch Intern Med 152:1299 – 1302.
Clinical Study
Correlation between early clinical response after catheter removal and diagnosis of catheter-related bloodstream infection Bassam Shukrallah, Hend Hanna, Ray Hachem, Dany Ghannam, Ioannis Chatzinikolaou, Issam Raad4 The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA Received 22 November 2006; accepted 16 March 2007
Abstract We conducted a retrospective post hoc analysis of prospectively collected data of cancer patients with central venous catheters (CVCs) who developed bacteremia with positive quantitative blood cultures (QBCs) drawn simultaneously through peripheral vein and CVC and which grew the same microorganisms from both blood cultures. We investigated whether clinical response of bacteremia, within 24, 48, or 72 h post-CVC removal, could be diagnostic of catheter-related bloodstream infection (CRBSI) when compared with microbiologic methods. Clinical response to antimicrobial therapy within 24 h of CVC removal in a patient with bacteremia was found to be highly suggestive of CRBSI, a finding that correlated well with semiquantitative catheter cultures and differential QBCs. However, response to antimicrobial therapy at z 48 h after CVC removal was less likely to be diagnostic of CRBSI and could reflect a response to antimicrobial therapy irrespective of the source of the bloodstream infections. D 2007 Elsevier Inc. All rights reserved. Keywords: Catheter removal; Bloodstream infections; Diagnosis of infections
1. Introduction The high frequency of catheter-related bloodstream infection (CRBSI), between 200 and 400 000 episodes occurring annually in the United States, is a public heath challenge ([Raad, 1998; Maki and Mermel, 1998; Raad and Hanna, 2002]). Not only do these infections present a significant monetary burden on the current healthcare system, where the estimated cost of managing each episode can range from $34 000 to $56 000 (Rello et al., 2000; Pittet et al., 1994), but they also are associated with a mortality rate of 12% to 35% (Collingnon, 1994; Pittet et al., 1994; Klugger and Maki, 1999). Accurately diagnosing CRBSI is essential for the successful management of such infections. Historically, clinicians arriving at the right diagnosis relied
4 Corresponding author. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA. Tel.: +1-713-792-7943; fax: +1713-792-8233. E-mail address: [email protected] (I. Raad). 0732-8893/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2007.03.012
on catheter removal and culture. However, only 4% to 10% of central venous catheters (CVCs) removed due to a suspected CRBSI were found to be colonized (Widmer et al., 1992). To address this problem, we developed newer diagnostic methods, such as differential time to positivity (DTP) and quantitative blood cultures (QBC) (Raad et al., 2004; Chatzinikolaou et al., 2006). Both of these methods provide sensitive and specific diagnostic capabilities without requiring the removal of the catheter. The Centers for Disease Control and Prevention (CDC) and Infectious Diseases Society of America (IDSA) recently incorporated these tests into their guidelines for the prevention of intravascular catheter-related infections (Mermel et al., 2001; O’Grady et al., 2002). Clinical response, defined as resolution of fever and other clinical manifestations of infection after removal of CVC, has also been considered in the absence of laboratory confirmation, as an indirect evidence of CRBSI (Pearson, 1996). This method has not been studied well, nor has it been compared with the established microbiologic methods. Furthermore, it lacks temporal parameters in the literature. The purpose of this study was to determine whether clinical
454
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457
Table 1 Clinical characteristics of patients with CRBSI and non-CRBSI Characteristic, n (%), unless indicated otherwise
CRBSI, n = 70
Non-CRBSI, n = 43
P
Sex, male Mean age F SD (years) Underlying malignancy Leukemia Lymphoma/myeloma Solid tumor ICU CVC type Silicone Rigid wall Port Quinton Neutropenia ( b 500 cells/mm3) Thrombocytopenia ( b 100 103/AL)
45 (64.3) 51.6 F 15.6
30 (69.8) 53.0 F 14.4
.55 .65
11 (15.7) 16 (22.9) 43 (61.4) 14 (20.0)
19 8 16 7
(44.2) (18.6) (37.2) (16.3)
.001 .59 .01 .62
62 3 4 1 22
(88.6) (4.3) (5.7) (1.4) (31.4)
29 (67.4) 11 (25.6) 2 (4.7) 1 (2.3) 26 (60.5)
.006 .001 .59 .62 .002
37 (52.9)
30 (69.8)
.08
response, within 24, 48, or 72 h of CVC removal, is indeed a useful diagnostic tool for CRBSI. Our current study focused on whether clinical response within 24 to 72 h, post-catheter removal, offers any diagnostic applicability when compared with semiquantitative catheter cultures (SQCCs) and QBCs. 2. Methods Between September 1999 and November 2000, we prospectively identified 113 cancer patients who, while having CVC in place, developed fever and had positive quantitative cultures of blood that was drawn simultaneously through peripheral vein and through the CVC and which grew the same microorganisms from both blood cultures. Patients were included in the study if their catheters were removed and cultured and if they did not have a complicated or disseminated bloodstream infection. 2.1. Definitions Catheter-related bloodstream infection. This is defined as the 1) fever and other clinical manifestations of bloodstream infection in a patient with CVC and 2) the presence of positive simultaneous QBC from the CVC and peripheral vein yielding the same organism, with 3) no apparent source for the bloodstream infection other than the catheter and 4) the presence of SQCC with significant catheter tip colonization with z 15 CFU of the same organism (same species and antibiogram) isolated from the blood cultures, and/or simultaneous QBC whereby the number of colonies isolated from the blood drawn through the CVC are at least 5-fold greater ( z 5:1) than the number of colonies isolated from blood drawn via a peripheral vein. CVC colonization. Significant colonization of the catheter tip was defined as a positive SQCC using the roll plate
method whereby at least 15 colonies of an organism was cultured from the catheter tip. Clinical response. This is the resolution of fever and other clinical manifestations of infection. 2.2. Culture techniques 2.2.1. Blood cultures After antiseptic cleansing of the skin and catheter hub, simultaneous quantitative and qualitative blood cultures were performed on blood drawn through the peripheral vein and the CVC’s hub. Blood samples drawn for cultures were handled as follows: 7 to 10 mL of blood was drawn through the CVC and discarded to avoid mixing with antibiotics or other agents that could have antimicrobial activity and that may have been given previously through the CVC. Subsequently, 20 mL of blood were drawn through the CVC and divided into 2 portions. Ten milliliters was placed in isolator tubes (Isolator 10; Wampole, Cranbury, NJ) to be cultured quantitatively according to the lysis centrifugation method as previously described (Dorn et al., 1979). Another 10 mL of blood was placed in a regular aerobic blood culture bottle (Aerobic 26+; BD Diagnostics, Sparks, MD). 2.2.2. Catheter culture Catheters were removed aseptically, and a 5-cm segment of the removed catheter tip was cultured using the semiquantitative roll plate method (Maki et al., 1977). 2.3. Statistical analysis Patients with CRBSI were compared with those with non-CRBSIs. The significance of the differences between the 2 groups was determined through the v 2 test or the Fisher’s exact test, as appropriate for categoric variables.
Table 2 Microbial organisms causing bloodstream infections in 113 cancer patients Organismsa
CRBSI (n = 70)
Non-CRBSI (n = 43)
P
Gram-positive organisms Coagulase-negative staphylococci Staphylococcus aureus a-Hemolytic Streptococcus Enterococcus spp. Other Gram-positive organismsb Gram-negative organisms Pseudomonas aeruginosa Escherichia coli Klebsiella pneumoniae Other Gram-negative organismsc Candida spp. Acid fast bacilli
51 32 9 3 2 5 13 3 2 5 3 5 1
30 (69.8%) 10 (23.3%) 12 (27.9%) 5 (11.6%) 2 (4.7%) 1 (2.3%) 11 (25.6%) 2 (4.7%) 4 (9.3%) 1 (2.3%) 4 (9.3%) 2 (4.7%) 0
.83 .018 .079 .26 .63 .4 .48 N .99 .2 .4 .42 .71 N .99
(73%) (45.7%) (12.9%) (4.3%) (2.9%) (7.1%) (19%) (4.3%) (2.9%) (7.1%) (4.3%) (7.1%) (1.4%)
a Numbers in parentheses are percentages of the total number of each isolated species. b h-Hemolytic Streptococcus, Bacillus spp., Corynebacterium, Micrococcus, and Stomatococcus. c Acinetobacter, Alcaligenes, Citrobacter, Enterobacter, Flavimonas, and Stenotrophomonas.
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457 Table 3 Clinical response after catheter removal Clinical response
CRBSI (n = 70)
Non-CRBSI (n = 43)
Within 24 h of CVC removal, n (%) Within 48 h of CVC removal, n (%) Within 72 h of CVC removal, n (%)
27 (37.1) 31 (42.9) 34 (47.1)
9 (23.3) 15 (34.9) 16 (37.2)
P .05 .323 .238
Student’s t test or Mann–Whitney U test was used for continuous variables. All P values were based on 2-tailed tests of significance with a level of significance at P V 0.05. Statistical analyses were performed using the statistical computing package SPSS (version 12.0 for Windows; SPSS, Chicago, IL). 3. Results Between September 1999 and November 2000, there were a total of 191 patients who had CVCs, developed fever, and had QBC done. The study included 113 of 191 patients who had their CVCs removed and cultured and who did not have complicated or disseminated infections. According to the definitions used in the study, there were 70 patients who had CRBSI and 43 who had non-CRBSI. The 2 groups were comparable in age, sex, and intensive care unit (ICU) admission status. Patients with CRBSI had solid tumor more often as their underlying malignancy ( P = 0.012), were less often neutropenic ( P = 0.002) and tended to be less thrombocytopenic ( P = 0.08), and had silicone CVCs more often (Table 1). Table 2 shows the distribution of microbial organisms causing CRBSI and non-CRBSI. The 2 groups were comparable in the distribution of major categories of infecting organisms (Gram-positive organisms, Gram-negative organisms, Candida spp., and acid fast bacilli). However, the frequency of coagulase-negative staphylococci causing CRBSI was significantly higher than those causing non-CRBSI (45.7% versus 23.3%, P = 0.016). Patients with CRBSI responded more often within the first 24 h after CVC removal ( P = 0.05) than did patients with non-CRBSI. This difference between the responses of the 2 groups disappeared at 48 and at 72 h post-CVC removal, when clinical response (CR) in patients with CRBSI was comparable to that of patients with non-CRBSI (Table 3). 4. Discussion This study shows that clinical response to antimicrobial therapy occurring within 24 h after the removal of the CVC in a patient with bloodstream infection may be suggestive of CRBSI. This suggested that clinical diagnostic criterion correlates well with microbiologic methods involving SQCCs and differential QBCs. However, response to antimicrobial therapy within 48 h or longer after CVC removal is less likely to be diagnostic of CRBSI and could
455
reflect the activity of the antimicrobial agents against the microorganism causing the bloodstream infections, irrespective of its source. The clinical diagnosis of CRBSI has been a challenge. On their own, clinical signs and symptoms are not useful in establishing the diagnosis of CRBSI because of their limited sensitivity or specificity (Mermel et al., 2001). The occurrence of fever or positive blood cultures in a patient with an indwelling CVC might not necessarily suggest, on its own, that the catheter is the source of the bloodstream infection. On the other hand, purulent discharge or inflammation involving the catheter exit site, the tunnel, or the pocket of the port might suggest that the catheter is a source of the bloodstream infection. However, although such findings are highly specific, they are not always associated with isolation of organisms that commonly cause CRBSI such as Staphylococcus epidermidis or Candida spp. Furthermore, evaluation of suspected infection in the presence of acute neutropenia is especially difficult because typical signs of inflammation may be markedly masked by the lack of neutrophils; hence, it is less likely that these patients would develop purulence at an infected catheter exit site. Because clinical findings are often associated with either poor sensitivity or poor specificity in establishing the diagnosis of CRBSI, the IDSA has recommended establishing the diagnosis based on microbiologic methods that include either quantitative catheter cultures or other differential blood culture methods, such as differential QBC or DTP (Raad et al., 2004; Chatzinikolaou et al., 2006; Mermel et al., 2001). Several studies have supported the use of such microbiologic methods (Raad et al., 2004; Chatzinikolaou et al., 2006; Flynn et al., 1988; Safdar et al., 2005; Blot et al., 1999; Gaur et al., 2003). In a recent metaanalysis, differential QBCs used in establishing the diagnosis of CRBSI in the study were considered the most highly specific and sensitive methods in making the diagnosis of such infection (Safdar et al., 2005). However, differential QBCs are not readily available in most hospitals in the United States or worldwide, and they are labor intensive and expensive. Furthermore, semiquantitative and quantitative catheter cultures are also costly and require the removal of the catheter. In their 1996 guidelines on the prevention of intravascular catheter-related infections, the CDC suggested that, in the absence of laboratory confirmation, defervescence after the removal of the suspected CVC from a patient with a bloodstream infection may be considered an indirect evidence of CRBSI (Pearson, 1996). Other investigators and reviewers have also relied on such clinical criteria (Blot et al., 1999). However, such criteria have not been well defined or tested against strict microbiologic methods that define CRBSI. Furthermore, neither the CDC nor the other investigators who referred to such clinical criteria outlined the time parameters necessary in establishing the diagnosis of CRBSI (24, 48, 72, or 96 h after CVC removal).
456
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457
The data in this current study suggest that a clinical response represented by defervescence after early CVC removal (within 24 h of such intervention) in patients with a bloodstream infection who are on appropriate antimicrobial therapy is suggestive of CRBSI. In our study, such early response does correlate well with the microbiologic definition of CRBSI, based on SQCCs and differential QBC as defined by the IDSA (Mermel et al., 2001). However, a clinical response occurring after more than 24 h of CVC removal does not suggest the diagnosis of CRBSI and does not correlate with the well-defined microbiologic criteria of CRBSI. The fact that the early response is indicative of the microbiologic diagnosis of CRBSI, rather than a later response, in a patient population on an appropriate antimicrobial therapy is quite revealing. Patients with uncomplicated bloodstream infections usually defervesce after at least 48 to 72 h of appropriate antimicrobial therapy (Fowler et al., 2003), irrespective of the source of the infection (catheter or otherwise). This occurs in cases of uncomplicated bloodstream infection in the absence of a deep-seated infection such as endocarditis or septic thrombosis. Patients with a definite uncomplicated CRBSI are more likely to respond earlier if the catheter source of the bloodstream infection is removed. Hence, the first 24 h are crucial in making the distinction between CRBSI and non-CRBSI cases. After 24 h, and particularly after 48 h of antimicrobial therapy, all patients with uncomplicated bloodstream infections would equally respond. This study, however, has several limitations. One of the major limitations is the small number of patients involved that included 113 patients who received appropriate antimicrobial therapy with an agent that is active against the organism isolated from the blood. The impact of CVC removal within 24 h was evaluated in a total of 36 patients, and the impact within 48 h was evaluated in only 48 patients. Hence, the lack of a difference, particularly in patients who defervesced within 48 or 72 h of CVC removal, could be related to the small numbers and the lack of power rather than the absence of impact of the CVC removal on establishing the diagnosis. Furthermore, the design of the study in terms of being a retrospective post hoc analysis of a prospective study represents another limitation. Finally, the fact that this study included cancer patients with different underlying malignancy and level of neutropenia might have affected the frequency of the early response in the CRBSI group. Patients with CRBSI were less likely to be neutropenic or to have underlying hematologic malignancy and, hence, they were more likely to respond earlier to antimicrobial therapy than the more immunocompromised neutropenic patients with underlying hematologic malignancy (Table 1). Finally, this method of establishing the clinical diagnosis requires the removal of the CVC; hence, in this current study, the impact of this method in establishing the diagnosis might have been limited to those cases of CRBSI
where the catheter was highly suspected and removed early during the bacteremia. Therefore, this method of clinically diagnosing the CRBSI might be useful retrospectively in surveillance rather than in establishing the diagnosis at the bedside. In addition, because this method requires the removal of the CVC, relying on it might lead to unnecessary and wasteful removal of the important intravascular device in the absence of other vascular access sites. In conclusion, based on a small number of patients, our data suggest that early clinical response to antimicrobial therapy, characterized by defervescence after 48 h of CVC removal, does not provide direct or indirect evidence suggesting the diagnosis of CRBSI. However, in a subset of patients who defervesce early within 24 h after the CVC is removed, the diagnosis of CRBSI is more likely and should be considered.
References Blot F, et al (1999) Diagnosis of catheter-related bacteraemia: a prospective comparison of the time to positivity of hub-blood versus peripheralblood cultures. Lancet 354:1071 – 1077. Chatzinikolaou I, et al (2006) Differential quantitative blood cultures for the diagnosis of catheter related bloodstream infections associated with short and long-term catheters: a prospective study. Diagn Microbiol Infect Dis 44:1834 – 1835. Collingnon PJ (1994) Intravascular catheter associated sepsis: a common problem. The Australian study on intravascular catheter associated sepsis. Med J Aust 161:374 – 378. Dorn GL, et al (1979) Improved blood culture technique based on centrifugation: clinical evaluation. J Clin Microbiol 9:391 – 396. Flynn PM, et al (1988) Differential quantitation with a commercial blood culture tube for diagnosis of catheter-related infection. J Clin Microbiol 26:1045 – 1046. Fowlerr Jr VG, et al (2003) Clinical identifiers of complicated Staphylococcus aureus bacteremia. Arch Intern Med 163:2066 – 2071. Gaur AH, et al (2003) Difference in time to detection: a simple method to differentiate catheter-related from non–catheter-related bloodstream infection in immunocompromised pediatric patients. Clin Infect Dis 37:469 – 475. Klugger DM, Maki DG (1999) The relative risk of intravascular device related infections in adults [Abstract]. Abstracts of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Francisco (CA)7 American Society for Microbiology, pp 514. Maki DG, et al (1977) A semiquantitative culture method for identifying intravenous-catheter–related infection. N Engl J Med 296:1305 – 1309. Maki DG, Mermel LA (1998) Infections due to infusion therapy. In Hospital infections. 4th ed. Eds, JV Bennett and PS Brachman. Philadelphia7 Lippincott-Raven, pp 689 – 724. Mermel LA, et al (2001) Guidelines for the management of intravascular catheter-related infections. Clin Infect Dis 32:1249 – 1272. O’Grady NP, et al (2002) Guidelines for the prevention of intravascular catheter-related infections. MMWR Recomm Rep 51:1 – 29. Pearson ML (1996) Guideline for prevention of intravascular device-related infections. Part I: Intravascular device-related infections: an overview. The Hospital Infection Control Practices Advisory Committee. Am J Infect Control 24:262 – 277. Pittet D, et al (1994) Nosocomial bloodstream infections in critically ill patients. Excess length of stay, extra costs, attributable mortality. JAMA 271:1598 – 1601.
B. Shukrallah et al. / Diagnostic Microbiology and Infectious Disease 58 (2007) 453 – 457 Raad I (1998) Intravascular-catheter–related infections. Lancet 351: 893 – 898. Raad II, Hanna HA (2002) Intravascular catheter-related infections. New horizons and recent advances. Arch Intern Med 162:871 – 878. Raad I, et al (2004) Differential time to positivity: a useful method for diagnosing catheter-related bloodstream infections. Ann Intern Med 140:18 – 25.
457
Rello J, et al (2000) Evaluation of outcome of the intravenous catheterrelated infections in critically ill patients. Am J Respir Crit Care Med 162:1027 – 1030. Safdar N, et al (2005) Meta-analysis: methods for diagnosing intravascular device-related bloodstream infection. Ann Intern Med 42:451 – 466. Widmer AF, et al (1992) The clinical impact of culturing central venous catheters. A prospective study. Arch Intern Med 152:1299 – 1302.